Re: kernel crash in mknod

[Roberto Sassu]

On 3/26/2024 12:40 PM, Christian Brauner wrote:
> > we can change the parameter of security_path_post_mknod() from
> > dentry to inode?
>
> If all current callers only operate on the inode then it seems the best
> to only pass the inode. If there's some reason someone later needs a
> dentry the hook can always be changed.

Ok, so the crash is likely caused by:

void security_path_post_mknod(struct mnt_idmap *idmap, struct dentry 
*dentry)
{
        if (unlikely(IS_PRIVATE(d_backing_inode(dentry))))

I guess we can also simply check if there is an inode attached to the 
dentry, to minimize the changes. I can do both.

More technical question, do I need to do extra checks on the dentry 
before calling security_path_post_mknod()?

Thanks

Roberto

> For bigger changes it's also worthwhile if the object that's passed down
> into the hook-based LSM layer is as specific as possible. If someone
> does a change that affects lifetime rules of mounts then any hook that
> takes a struct path argument that's unused means going through each LSM
> that implements the hook only to find out it's not actually used.
> Similar for dentry vs inode imho.